Escolar Documentos
Profissional Documentos
Cultura Documentos
Japan Aerospace Exploration Agency, Tsukuba Space Center, 2-1-1 Sengen, Tsukuba, Ibaraki 305-8505, Japan
INO, Remote Sensing Group, 2740 Einstein, Quebec, Quebec G1P 4S4, Canada
Korea Research Institute of Standards and Science, 267 Gajeong-ro, Yuseong-gu, Daejeon 305-340, Republic of Korea
d
Institut fur Materialphysik im Weltraum, Deutsches Zentrum fur Luft- und Raumfahrt (DLR), 51170 Koln, Germany
e
California Institute of Technology, Materials Science Department, Mail Stop 138-78, Pasadena, CA 91125, United States
f
Japan Science and Technology Agency, Precursory Research for Embryonic Science and Technology (PRESTO), Kawaguchi, Saitama 332-0012, Japan
g
Department of Science of Measurement, University of Science and Technology, Daejon, 305-333, Republic of Korea
b
c
A R T I C L E I N F O
Article history:
Available online xxx
Keywords:
Electrostatic levitation
Materials processing
Supercooling
Nucleation
Thermophysical properties measurement
Atomic dynamics
A B S T R A C T
Electrostatic levitators have been around for more than 30 years and have become mature tools for the
material science community. Originally developed as positioners for materials and uid science
experiments in space, they saw a myriad of offsprings throughout the world for ground-based research,
not only in space agencies but also in governmental laboratories, in universities and in the industry.
Electrostatic levitators eliminate any physical contact with a container allowing to process and study
corrosive or high temperature materials in their solid or liquid phases. Moreover, heterogeneous
contamination from the container being avoided, it is possible to reach and maintain supercooled and
metastable phases. This, in turns, permits a host of fundamental and applied studies. The nucleation and
solidication phenomena can be scrutinized, the atomic structure and dynamic of liquid and metastable
phases can be probed and the physics of molten drops could be investigated. On a more applied
standpoint, the measure of thermophysical properties and the synthesis of materials with new properties
are also possible with current facilities.
This paper rst describes the principle of electrostatic levitation and retraces the development of
various facilities throughout the world, focusing on the advances made by each research group. The
capabilities of electrostatic levitation for materials processing and synthesis under different
environments are then presented. The paper successively covers in length its contribution for the
measurements of thermophysical properties and for fundamental studies using high energy particle
beams. Finally, the outlook of electrostatic levitators and its attractiveness for space experiments in
materials sciences are discussed.
2013 Elsevier B.V. All rights reserved.
Abbreviations: AC, alternative current; A/D, analog to digital; ADL, aerodynamic levitation; BESL, beam electrostatic levitator; BMG, bulk metallic glass; Caltech, California
Institute of Technology; CCD, charge coupled device; CN, coordination number; CNT, classical nucleation theory; CPMD, Car-Parrinello molecular dynamics; D/A, digital to
analog; DC, direct current; DISRO, distorted icosahedral short-range ordering; DLR, Deutsches Zentrum fur Luft- und Raumfahrt (German Aerospace Center); DSC, differential
scanning calorimetry; DTA, differential thermal analysis; ESCAPES, electrostatic containerless processing system; ELF, electrostatic levitation furnace; ESA, European Space
Agency; ESL, electrostatic levitator; FTIR, Fourier transform infrared; GFA, glass forming ability; HDL, high density metallic liquid; Hi-Lite, High Temperature Levitation
Instruments for Thermophysical Evaluation; HRIS, high resolution inelastic scattering; HRPD, high resolution power diffractometer; HTESL, high temperature electrostatic
levitator; IHI, Ishikawajima Heavy Industries; ISRO, icosahedral short-range ordering; ISS, International Space Station; ITS, International Temperature Scale; JAERI, Japan
Atomic Energy Research Institute; JAXA, Japan Aerospace Exploration Agency; JPL, Jet Propulsion Laboratory; KRISS, Korea Research Institute of Standards and Science; LDL,
low density semimetallic liquid; LLPT, liquid liquid phase transition; MCPD, multi channel photo detector; MLWF, maximally localized Wannier functions; MRO, medium
range order; NASA, National Aeronautics and Space Administration; NASDA, National Space Development Agency; NIMS, National Institute for Materials Science; MSFC,
Marshall Space Flight Center; PID, proportional integral derivative; PSD, position sensing detector; RF, radio-frequency; RMC, reverse Monte-Carlo calculations; SEM,
scanning electron microscope; SRO, short range order; TTT, temperature time transformation; UV, ultraviolet; WU-BESL, Washington University Beam Electrostatic Levitator;
YAG, yttrium aluminum garnet.
* Corresponding author at: INO, Remote Sensing Group, 2740 Einstein, Quebec, Quebec G1P 4S4, Canada.
E-mail address: paul-francois.paradis@ino.ca (P.-F. Paradis).
0927-796X/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.mser.2013.12.001
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Contents
1.
2.
3.
4.
5.
6.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Electrostatic levitation technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Development and description of facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.
Battelle Institut . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.1.
National Aeronautics and Space Administration-Jet Propulsion Laboratory .
2.1.2.
2.1.3.
Space Systems Loral/NASA-Marshall Space Flight Center/Washington U . . .
National Space Development Agency/Japan Aerospace Exploration Agency .
2.1.4.
National Institute for Materials Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.5.
German Aerospace Center (DLR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.6.
2.1.7.
Korea Research Institute of Standards and Science . . . . . . . . . . . . . . . . . . . .
Northwestern Polytechnical University . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.8.
Charging strategy and levitation initiation [6,10] . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.
Capacitive charging [6,10] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1.
2.2.2.
Photoelectric emission charging [6,10] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Thermionic emission charging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.3.
Heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.
Sample rotation control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.
2.4.1.
Importance of rotation control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Rotation control methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.2.
Materials processing and synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Crystallization and vitrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.
3.2.
Solidication studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nucleation studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.
3.4.
Phase diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reduced-gravity processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.
Materials thermophysical property measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Density and thermal expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.
Measurement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1.
Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.2.
Surface tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.
Measurement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.1.
Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.2.
Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.
Measurement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.1.
Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.2.
Ratio of heat capacity to emissivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.
Measurement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.1.
Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.2.
Emissivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.
Measurement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.1.
Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.2.
Electrical conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.
Measurement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.1.
Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.2.
Vapor pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.
Measurement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.1.
Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.2.
Short-range structure and atomic dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hard sphere model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.
Neutron scattering experiments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.
Preliminary experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.1.
Elastic and quasi-elastic experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2.
Synchrotron scattering experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.
Inelastic scattering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.1.
Elastic scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.2.
Conclusions and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction
The processing of materials without any containers, crucibles,
or walls has several advantages [13]. For one, it allows the
processing of corrosive materials or materials whose melting
temperatures are higher than that of the crucible. Handling of high
temperature materials in their solid or liquid phases is thus
achievable. It also eliminates or reduces the temperature gradients
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
Table 1
Comparative summary of the characteristics of different levitation techniques with regards to materials processing and study.
Levitation
technique
Processing environment
Material
Access to
sample for
diagnostics
Stability at high
temperature
Comments
Type
Size
High vacuum
Pressurized
atmosphere
Electrostatic
Conductor;
semi-conductor;
insulator
mm-size
Possible
Excellent
Excellent
Electromagnetic
Conductor
cm-size
Possible
Possible but
pressure range
limited by
Paschens rule
Possible
Fair
Good
Acoustic
Conductor;
semi-conductor;
insulator
Conductor;
semi-conductor;
insulator
Conductor;
semi-conductor;
insulator
Conductor;
semi-conductor;
insulator
cm-size
Impossible
Possible
Good
Bad
mm-size
Impossible
Possible
Fair
Good
cm-size
Impossible
Possible
Excellent
Good
Atoms to
mm-size
objects
Possible
Possible
Excellent
Excellent
Aerodynamic
Ultrasonic
Optical
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
e1
:
e1
(1)
1
rc U 2
2
(2)
Fig. 1. Pictorial representation of the electrode conguration of JPL early electrostatic levitators: (a) dish type; (b) ring type; (c) tetrahedral type [6].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
Fig. 2. Block diagram of the main position control components of the JPL early
electrostatic levitators (shown here for the tetrahedral positioner conguration)
[6].
Fig. 3. Schematic representation of the electrode conguration and active position control hardware used by the JPL high temperature electrostatic levitator (shown in one
plane).
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
Fig. 5. Schematic representation of the NASDA electrostatic levitation dedicated to the sounding rocket TR-1A [16].
Reprinted with permission from S. Yoda et al., J. Japan Soc. Microg. Appl. 17, 7686. Copyright 2000, Japan Society of Microgravity Application.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 7. Laser beams conguration in JAXA vacuum facility to heat a levitated sample.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
Fig. 8. Pictorial representation of the main system components (view from the top) of the JAXA vacuum electrostatic levitator.
Fig. 9. Pictorial representation of the main system components (view from the top) of the JAXA pressurized electrostatic levitator.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 11. Illustration of the processing chamber (left) and of the sample cartridge
(right) [62].
Reprinted with permission from N. Sugi et al., J. Jpn. Soc. Microg. Appl. 23, 2025.
Copyright 2006, Japan Society of Microgravity Application.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
11
Table 2
Specications of the JAXA ISS electrostatic levitation furnace [62,63].
Item
Requirement
Sample diameter
Position stability
Atmosphere
25 mm
100 mm
Ar/N2 up to 2 atm
Low pressure (103 Torr) through
waste gas line
4 diode lasers (970 nm) (200 W total)
Pyrometry
1 for density measurement and
1 for general view
Density by image analysis
Surface tension and viscosity
by oscillation drop method
>10,000 frame per second
Heating
Temperature measurement
Camera
Thermophysical property
measurement
High speed camera
Fig. 12. Electrode conguration of the breadboard model for JAXA ISS electrostatic
levitation furnace [62].
Reprinted with permission from N. Sugi et al., J. Jpn. Soc. Microg. Appl. 23, 2025.
Copyright 2006, Japan Society of Microgravity Application.
Fig. 13. Photograph of the breadboard model for JAXA ISS electrostatic levitation
furnace [62].
Reprinted with permission from N. Sugi et al., J. Jpn. Soc. Microg. Appl. 23, 2025.
Copyright 2006, Japan Society of Microgravity Application.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
ceramic rod that can be moved through a hole in the center of the
bottom electrode, thus raising the solid sample to a position at which
it can be illuminated and heated by the heating lasers.
To compensate the charge losses during heating of the sample
before charging by thermionic emission becomes efcient at
elevated temperatures, the specimen is illuminated by the
ultraviolet light of a deuterium lamp. This helps the specimen
to gain charge by means of the photoelectric effect.
The stationary electrostatic levitators at DLR are utilized to
study the supercooling behavior of various melts of metals and
semiconductors [68], to measure the thermophysical properties of
the melts such as density and viscosity [69], and to investigate the
crystal growth behavior during non-equilibrium solidication of
supercooled melts [70,71]. The design of electrostatic levitators
described above is also the basis of DLRs mobile ESL which is used
for some of the scattering experiments that are described in
Section 5. Moreover, the sample position control technology
developed at DLR constitutes the heart of Washington Universitys
WU-BESL [40] and of a facility at Iowa State University [31].
2.1.7. Korea Research Institute of Standards and Science
In the view of standards in metrology, levitation techniques are
attractive. Up to date, the denition of the International
Temperature Scale (ITS) was based on a series of metal xed
points from which freezing temperatures are dened as closer to
their thermodynamic temperatures as is possible. To obtain the
metal xed points, a crucible that contains molten metals is
needed. However, the crucible can be a source of impurities, and
react with the molten metals causing contamination. Thus, the
crucible can affect the measurement of the xed point temperatures. Moreover, mechanical stress after repeated melting and
freezing processes can causes the breakage of even graphite
crucibles [72]. Although the metal-carbon eutectics may be
accepted as a clever way to circumvent the contamination of
metals and breakage of the container, this method cannot avoid the
reaction with the container wall. Furthermore, a temperature
gradient between a heat source and the sample, and a time lag of
the freezing temperature recovery after supercooling hinders
accurate measurement. Therefore, it is highly desired that the
sample should be isolated from the container so that thermal
radiation from the samples alone could be studied.
Levitation techniques, in particular, electrostatic levitation
operating under high vacuum, can provide the desired solution.
The containerless environment given by the ESL allows controlling
the amount of supercooling of the molten sample by adjusting the
power of heating lasers. Unlike the contact method using a
crucible, the ESL technique offers a clear physical meaning of
supercooling, recalescence, and plateau which subsequently
accompany the release of latent heat of crystallization, and are
useful information to determine the standard temperature. In
addition, in a metrology standards view point, the secondary
thermophysical properties of molten material liquids at high
temperature are doubtful. Therefore more reliable thermophysical
properties (e.g., heat capacity, density, viscosity, surface tension,
emissivity, thermal conductivity, and fusion enthalpy) of molten
liquids can be obtained using ESL compared to other levitation
techniques. A vacuum ESL does not have any contamination from
the presence of gas, neither the convection induced either by an
induced current or a gas ow. The shape of the molten sample
being very close to that of a sphere, no complex analysis of the
measured quantities is necessary. For these specic reasons, the
Korea Research Institute of Standards and Science (KRISS) decided
to develop an electrostatic levitator, called Hi-Lite (High Temperature Levitation Instruments for Thermophysical Evaluation), for
the measurement of thermophysical properties for standards as
well as materials science at high temperatures.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
13
(3)
Fig. 17. Laser beams conguration in the JAXA vacuum facility for sample levitation
initiation by thermionic effect and during levitation.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
a 2p f r
(4)
where f is the rotation frequency and r the sample radius. From this
relationship, the radial acceleration normalized with the gravitational acceleration (g-level) at sea level as a function of rotation
frequency for samples of different diameters can be found. It
indicates that millimeter-size samples with a rotation rate as low
as 1 Hz would experience g-levels well over those of the
International Space Station [79]. Therefore the advantages of
microgravity environment would be completely spoiled. Experimental results presented by Professor R.K. Wunderlich (Ulm
University) at a seminar sponsored by the Japan Space Forum in
Tokyo, Japan, in 2002 dramatically showed that under certain
circumstances, electromagnetically levitated samples can be
deformed due to rotation occurring in microgravity. The reasons
of sample rotation, on the ground as well as in microgravity, are
multiple and depend on the levitation and heating schemes. In the
present case, a possible source of rotation is the slight misalignments of the heating laser beams. Therefore, the use of a rotation
control device is necessary not only for ground-based experimental work but also to take full advantage of the reduced gravity
conditions offered by an orbital platform such as the ISS.
2.4.2. Rotation control methods
Different approaches can be used to control rotation of levitated
samples. Their application depends on the processed material
(electrical conductor, insulator), the rapidity of desired implementation, the accuracy, and the processing environment (high
vacuum, pressurized atmosphere, rareed gas, etc.). Four schemes,
namely the rotating magnetic eld, the linear and angular photon
momentum transfer, and the ultrasonic pressure, are shortly
reviewed and discussed next.
2.4.2.1. Rotating magnetic eld. For electrical conductors, the use of
rotating magnetic elds is very attractive. Following earlier work
[27], four coils are used as a stator to generate a horizontal and
rotating magnetic eld (Fig. 17). Each of the coils is wound on a
glass spool with an insulated copper wire. The glass is used to
electrically insulate the spools from the proximity of the high
voltage electrode. A soft iron core is inserted in the spool to
increase the magnetic eld [27]. In JAXA vacuum facility, the coils
surrounded the lower electrode [80] to ease maintenance and to
lower the risk of arcing between the high voltage top electrode and
the coils. Plates of MacorTM further prevented any arcing and
protected the coil assembly from contact with an unsuccessfully
processed molten sample (Fig. 18).
The principle of this technique relies on that of the asynchronous induction motor, where a stator generates the magnetic eld
and where the rotor is the sample. If an ac voltage V at frequency vv
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
15
Fig. 18. Side view of the rotation coils as implemented on the JPL high temperature
electrostatic levitator. A charged conducting sample can be rotated around the
vertical direction when a rotating eld is applied in the same direction [27].
tM
vv V 2
R2ST
2 2
v LST
!
1
RSP
1v
vv
(5)
tL
4pr 2 LeT s SB T 4
c
(6)
where RSP and v are respectively the resistance and the rotation
frequency of the sample.
To generate the magnetic eld, a signal generator sent two
sinusoidal signals with a p/2 phase difference. Each of these signals
was amplied and sent to a control box that allowed the choice
between horizontal and vertical elds. The box also permitted to
select a clockwise or anti-clockwise eld rotation (Fig. 19).
As was experimentally determined [27], the applied torque
necessary to spin a 2.38 mm diameter sample at 200 Hz is around
2 1011 N m when the rotation frequency of the applied
magnetic eld is 400 Hz and the coil current is 0.2 A.
Accurate measurement of sample rotation frequency is important for accurate determination of torque induced to the levitated
sample. To measure the sample rotation frequency, the reection
of a low power (10 mW) HeNe laser beam impinging onto the
sample surface was detected by a silicon photodetector. The output
voltage of the detector was amplied and digitized to get a Fourier
power spectrum using a LabVIEWTM software package. Such a
power spectrum showed peaks at all harmonics of the sample
rotation frequency. In addition, the stroboscopic effects generated
by the television monitor helped to visually determine the
frequency at 30 Hz intervals.
Fig. 19. Block diagram of the electronic components that generate appropriate
electric currents to the coils to produce the desired magnetic eld necessary for
sample rotation [27].
Fig. 20. Photographs showing, from the side, a Zr sample being subject to a rotating
magnetic eld, from a static state up to bifurcation [77].
Reprinted with permission from P.-F. Paradis et al., J. Japan Soc. Microg. Appl. 20,
218225. Copyright 2003, Japan Society of Microgravity Application.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
with the electrostatic facilities around the world on millimetersize samples would require time scales too long that would hinder
practical applications. Detailed calculations for electrostatic
levitation applications are presented in the literature [77].
2.4.2.3. Ultrasonic pressure. As was shown by Trinh [83], the
rotation of acoustically levitated sample under a pressurized
atmosphere is possible. The origin of the torque, as evidenced by
ow visualization, appears to be linked with steady acoustic
streaming ows induced by high intensity sound waves [84]. This
was made possible through small horizontal translation adjustments made either to a reector or a transducer relative to their
axially symmetric position. Although no relationship was given
between power and sample size, the rotation rate could be
controlled by varying the ultrasonic level while keeping the
reector and transducer displacement from symmetrical in a xed
position.
3. Materials processing and synthesis
3.1. Crystallization and vitrication
Fig. 21. Photographs showing, from the side, a Zr sample being subject to a linear
photon momentum transfer, from a static state all the way to bifurcation [77].
Reprinted with permission from P.-F. Paradis et al., J. Japan Soc. Microg. Appl. 20,
218225. Copyright 2003, Japan Society of Microgravity Application.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
17
Fig. 24. Temperature as a function of time for the free cooling of liquid
Ti39.5Zr39.5Ni21 in ESL. Double recalescence on small supercooling (a), and single
recalescence for i-phase from deeply supercooled liquid (b) [90].
Reprinted with permission from G.W. Lee et al., J. Acta Mater. 59, 49644973.
Copyright 2011, Elsevier.
Fig. 22. Two-step recalescence for alloys containing 21 at.% Ni and different Ti to Zr
ratio [87]. The formation of the i-phase occurs in (b)(d).
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 72, 174107. Copyright
2005, American Physical Society.
Fig. 23. (a) Cooling curves for an electrostatically levitated Ti39.5Zr39.5Ni21 droplet
(2.5 mm diameter) as a function of temperature showing two recalescence events
(indicated by arrows). (b) X-ray diffraction patterns for the supercooled liquid
Ti39.5Zr39.5Ni21 at 1029 K, during the 1st recalescence to the i-phase, and during the
2nd recalescence to the C14 Laves phase [88].
Reprinted with permission from K.F. Kelton et al., Phys. Rev. Lett. 90, 195504.
Copyright 2003, American Physical Society.
fact from this result is the critical cooling rate for the formation of
the i-phase. As the supercooled liquid drops onto a Cu rod in the
ESL, the critical cooling rate of the i-phase formation for TiZrNi
alloys was estimated to be less than 100 K/s (Fig. 25) [90]. This
value is much smaller than that of other metastable quasicrystal
formation (>500 K/s) [91].
The critical cooling rate is particularly important in bulk
metallic glass (BMG) study, since it is a key parameter to determine
glass forming ability (GFA). For the determination of the critical
cooling rate, a temperature-time transformation (TTT) diagram
should be constructed. The construction of a TTT diagram is very
hard to achieve in conventional contact methods since it requires
that the supercooled liquid should be maintained at certain
temperatures. ESL has been used for the construction of such
diagrams for obtaining the critical cooling rates of BMG [9294].
For instance, Fig. 26 shows the TTT diagram for Pd40Cu30Ni10P20.
PdCuNiP, ZrTiNiCuBe (Vit 1), and ZrNbCuNiAl
exhibited critical cooling rates of 0.33 K/s [92], 2 K/s [93], and
1.5 K/s [94], respectively.
ESL raised a question for the GFA of BMG. In the conventional
contact methods, it is expected that the larger size BMG should
Fig. 25. Cooling curves of Ti37Zr42Ni21 after quenching from 758 8C onto a Cu rod.
The cooling rates for each region are shown [90].
Reprinted with permission from G.W. Lee et al., J Acta Mater. 59, 49644973.
Copyright 2011, Elsevier.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Temperature (K)
2500
Tm
2000
1500
1000
Time (s)
Fig. 28. Typical cooling curve for an Y3Al5O12 sample (12.61 mg) (120 Hz sampling
rate) [98].
Fig. 27. Free radiation cooling curves for levitated (a) Zr57Ti5Cu20Ni8Al10 and (b)
Zr62Cu20Ni8Al10 alloys from the liquid to the solid phases. (b) Does not show a
recalescence, signaling glass formation, while (a) shows a recalescence near 790 K,
indicating crystallization [95].
Reprinted with permission from T.H. Kim et al., Appl. Phys. Lett. 87, 251924.
Copyright 2005, AIP Publishing LLC.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
DG n
16p
s3
;
3 dm=v2
Is
6n 2=3 kB TNA
phTl
jdmj
6pkB Tn
1=2
DG n
exp
kB T
(8)
(b) 2800
2300
Tm
2200
2100
2000
1900
1800
2600
Tm
2400
- allotropic transition
2200
2000
1800
1600
1700
0 1 2 3 4 5 6 7 8 9 10111213
Time (s)
(c) 4000
Time (s)
(d) 4000
3500
Temprature (K)
Temperature (K)
(7)
where s, dm, and v are the interfacial free energy, the Gibbs free
energy difference between the initial and nal phases per atom, and
the average atomic volume v. The direct measurement of the
interfacial free energy is very difcult, since the interface is too small
to measure and because it is composed of only a few atomic layers.
Whenever a container is used to conne a liquid, the container
provides heterogeneous nucleation sites that the liquid cannot
deeply cool down. This complicates the analysis to obtain the
interfacial free energy using the classical homogeneous nucleation
theory (CNT). Electrostatic levitation can resolve the difculties,
since it provides a containerless environment. According to the CNT,
the homogeneous steady state nucleation rate is given by
Temperature (K)
Temperature (K)
(a) 2400
19
Tm
3000
2500
2000
Tm
3500
3000
2500
2000
1500
1500
4
6
Time (s)
10
Time (s)
Fig. 29. Thermal history of selected millimeter-size samples processed with JAXA vacuum levitation furnace: (a) Rh (Tm = 2236 K; 185 K/s average cooling rate, 184 K
supercooling) [104]; (b) Hf (Tm = 2504 K; 710 K/s average cooling rate, 285 K supercooling; ab allotropic transition) [105]; (c) Re (Tm = 3441 K; 1100 K/s average cooling
rate, 915 K supercooling) [106]; (d) W (Tm = 3695 K; 1820 K/s average cooling rate, 500 K supercooling) [48].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Table 3
Interfacial free energies, a = s/DHf, work for critical cluster formation, and critical radius r* for the three phases (i, Laves, and b).
Phases
i-phase (Ti39.5Zr39.5Ni21)
C14 Laves (Ti40Zr30Ni30)
Solid solution phase (Ti49Zr49Ni2)
s (J/m2)
0.062 0.002
0.095 0.006
0.077 0.002
aLS
W*/kBTr
0.324 0.009
0.375 0.017
0.59 0.013
58.28 0.15
60.68 0.29
61.21 0.40
where h, l, n*, kB, NA and DG n are the viscosity, the average atomic
jump distance, the number of atoms in the critical nucleus, the
Boltzmanns constant, and the Avogadro number, respectively. For
the initiation of the crystallization from a supercooled liquid, at
least one critical size nucleus should form. This requires that the
number of the critical nuclei in the supercooled liquid with a given
liquid volume V, nucleation rate Is, and time tr, should be greater
than one, as follows [125]:
Is T r V t r
1
(9)
17.23 0.32
14.38 0.57
19.56 0.13
Used parameters
r (g/cm3)
Cp (J/mol-K)
DHf (kJ/mol)
5.712
5.448
5.884
44.24 1.54
42.45 3.78
33.57 1.65
8.48 0.40
11.2 0.42
6.12 0.13
Fig. 31. (a) Calculated work for critical cluster formation and (b) critical radius for
nucleation of the three phases using Turnbulls driving free energy expression
(dm = DHf DT/Tm) [87].
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 72, 174107. Copyright
2005, American Physical Society.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
21
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 34. Temperature-time cooling curves for ESL-processed TiZrNi alloy liquids
for low Ni concentrations that nucleate b(Ti/Zr) as the primary crystallizing phase
[87].
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 72, 174107. Copyright
2005, American Physical Society.
Fig. 35. (a) X-ray diffraction for low Ni concentrations showing a-solid solution (&) and C14 Laves phase (O). (b) SEM microstructures of Ti49Zr49Ni2. (c) C14 Laves phase
forms on Ti47.5Zr47.5Ni5, which corresponds to the small plateau near 750 8C in Fig. 33 (b) [87].
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 72, 174107. Copyright 2005, American Physical Society.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
Fig. 36. Pseudo-binary phase diagram of TiZrNi alloys with equal ratio of Ti and
Zr. Maximum supercooling temperatures (+) are given from [87,90].
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 72, 174107. Copyright
2005, American Physical Society; Reprinted with permission from G.W. Lee et al., J
Acta Mater. 59, 49644973. Copyright 2011, Elsevier.
23
Fig. 38. Pictures taken from a video camera taken during a successful zero-gravity
launching of a compound drop using the JPL tetrahedral positioner. It shows one of
the rst successful positioning of a compound liquid drop in reduced gravity. (a) A
dyed water drop with an air core is attached to the nozzle and resting on a velvet
pad, ready to be launched. (b) It is repelled from the nozzle with the application of
voltage under servo control. (c) The neck is being broken showing a Taylor cone. (d)
The compound drop is in the air and being held at the pre-set position [6].
Reprinted with permission from W.-K. Rhim et al., Rev. Sci. Instrum. 56, 307317.
Copyright 1985, AIP Publishing LLC.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Li2O76 mass% SiO2 composition was doped with 0.1 m% of Co(II)oxide to increase its optical absorptivity [5,24]. The sample was
preheated to the transformation temperature prior to the launch.
Once in microgravity, it was released by a sample handling
mechanism and maintained in the focus of a heating mirror by the
electrostatic positioning system. By applying high voltages (from
3.5 to 5 kV) between the electrodes, the uncharged, polarized glass
sample was supposed be positioned. However, a misalignment of
one of the CCD camera used for position control occurred during
the launch. As a result, the electrodes were thus wrongly activated
and no proper positioning was achieved. The sample moved
irregularly between the electrodes and, therefore, being not always
in the mirror focus, it reached a temperature of only 500700 8C,
instead of the target 1150 8C, before it stuck to one of the
electrodes. The nuclei did not dissolve and phase separation
occurred. As shown in Fig. 39, the initially transparent blue sample
turned to blue opaque.
Containerless solidication of BiFeO3 has been carried out by
NASDA (JAXA) in microgravity with an electrostatic levitation
furnace on board a sounding rocket (TR-IA) [16]. The aim of this
experiment was to study the solidication behavior of oxide
insulator material in microgravity. A spherical BiFeO3 specimen
with a diameter of 5 mm was laser heated and solidied in an
oxygen and nitrogen mixture atmosphere. The microstructure
resulting from solidication in the levitation furnace was
compared with that obtained from solidication in a 10 m drop
tube and in crucibles [43]. In the crucible experiments, the
segregation of the primary Fe2O3 phase could not be suppressed,
even if the cooling speed increased to 5000 K/s. However it did
suppress in a 0.3 mm diameter droplet solidied in the drop tube
experiment. This suggests that containerless processing effectively
promoted the supercooling of the BiFeO3 phase although the
positioning system did not work as expected, therefore inducing
thermal gradient in the sample as it moved with respect to the
heating laser beams (Fig. 40). In the microgravity experiment,
although a homogeneous BiFeO3 phase was not observed in the
5 mm spherical specimen, an anomalous ne cellular microstructure appeared due to deep supercooling. In addition, the phase
transitions of BiFeO3 were measured by DTA from room
temperature to 1523 K and its liquidus temperature was estimated
to be 1423 K.
Following extensive ground-based research and development
in terms of position control system and thermophysical property
measurements, additional testing was carried out in reduced
gravity by JAXA with a dedicated facility [62]. Experiments were
done aboard the Mitsubishi MU-300 aircraft. During the parabolic
ight, stable positioning, laser heating and melting, cooling,
supercooling, and solidication of a Zr sample could be successfully completed.
Fig. 39. Photographs of the sample before (a) and after (b) processing in reduced
gravity.
Fig. 40. Laser power and temperature proles measured by pyrometry (a) and video
images illustrating the uctuating position of the sample during heating and
cooling while in reduced gravity (b) [16].
Reprinted with permission from S. Yoda et al., J. Japan Soc. Microg. Appl. 17, 7686.
Copyright 2000, Japan Society of Microgravity Application.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
25
Fig. 41. Schematic of DLR optical setup for the measurement of the density with an
electrostatic levitator.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 42. Typical measurements for the density obtained with electrostatic levitation coupled with an imaging technique: (a) Si [77].
Reprinted with permission from P.-F. Paradis et al., J. Japan Soc. Microg. Appl. 20, 218225. Copyright 2003, Japan Society of Microgravity Application; (b) W [48]: (c) Ir [152];
and (d) BaTiO3 [115].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
Tm (K)
Ce
La
Pr
Nd
Au
Gd
Tb
Si
Ni
Y
Pd
Ti
1072
1191
1204
1294
1337
1586
1629
1683
1728
1796
1828
1941
Pt
Zr
2041
2128
V
Rh
B
Hf
Ru
Ir
2183
2237
2360
2504
2607
2719
Nb
Mo
Ta
2750
2896
3290
Os
Re
3306
3441
3695
Density r
Reference
r(Tm)
(103 kg/m3)
dr/dT
(kg/m3 K1)
6.41
6.00
6.50
6.59
17.4
7.41
7.84
2.55
7.89
4.15
10.66
4.21
4.17
19.2
6.34
6.21
5.46
10.82
2.17
11.82
10.75
19.87
19.5
7.73
9.11
14.75
15.0
19.10
18.65
18.4
16.43
0.83
0.39
0.51
0.57
1.44
0.46
0.47
0.65
0.21
0.77
0.51
0.22
0.96
0.15
0.27
0.49
0.76
0.25
0.55
0.56
0.71
0.85
0.39
0.60
0.85
0.41
1.16
0.79
0.91
1.08
27
Table 5
Measured density data of molten oxide materials [97,98,115,116].
Composition
Tm (K)
Density r
3
[156]
[157]
[157]
[157]
[158]
[159]
[160]
[77,153]
[161]
[162]
[163]
[164]
[165]
[166,167]
[32]
[168]
[169]
[104]
[170]
[105]
[171]
[171]
[152]
[172]
[173]
[171]
[106]
[175]
[106]
[174]
[48,176,177]
important to point out that the entire data taking process is carried
out over a very short period (a few seconds). Not only does this
ensure that the impact of evaporation is minimized, but this also
permits several data to be acquired in a single day, thus giving the
opportunity to validate reproducibility. Except for the quadratic
behavior exhibited by Si (agrees with [153]), the linear dependence
with temperature of the metals compares well with the theoretical
works of Reynolds [154] and Steinberg [155]. Most of the data
measured so far with electrostatic levitation combined with the
imaging technique are summarized in Table 4 [32,48,104
106,77,153,156177].
At the melting point, the values agrees generally very well with
those with those measured by other techniques (e.g., drop weight
technique in vacuum, the isobaric expansion method, the pendent
drop method, the sessile drop method, and, in some case,
aerodynamic or electromagnetic levitation coupled with imaging)
when respective experimental uncertainties are taken into
account. The discrepancy observed between the results (values
at the melting temperature as well as temperature coefcients)
obtained with electrostatic levitation and other techniques could
be attributed to the difference in processing techniques and the
extent to which evaporation losses have been considered. The
containerless approach in high vacuum offered by electrostatic
levitation, isolated the samples from container walls and gases,
whereas the other methods often implies possible chemical
reactions between the highly reactive molten materials and a
crucible, a support, or residual gases. This can alter the nal density
values. Other possible sources of error could arise from the imaging
technique, in particular the focusing of the sample, the digitization
of the image, and the measurements of sample mass. Material
purity, oxygen solubility in the samples, gasication, surface
oxidation or nitridation and level of vacuum could also, to some
extent, explain the discrepancies. Temperature measurements too
BiFeO3
Nd-CA
BaTiO3
YAG
Al2O3
1363
1878
1893
2240
2327
Reference
3
dr/dT (kg/m K)
6.74
2.83
4.04
4.08
2.93
1.31
0.21
0.34
0.29
0.12
[115]
[97]
[115]
[98]
[116]
(12)
Fig. 43. Density of liquid Zr64Ni36 versus temperature measured by DLR using
electrostatic levitation [69,150]. The solid and hollow symbols, respectively,
correspond to two different runs with individually prepared samples. For
comparison, the measurement of Ohsaka is shown by the dashed line [179].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
28
Table 6
Parameters for the molar volume and density of liquid ZrNi binary alloys measured using the electrostatic levitation technique.
Alloy
TL (K)
VL (cm3/mol)
a (105/K)
r(TL) (g/cm3)
Reference
Ni50Zr50
Ni36Zr64
Ni36Zr64
Ni33.33Zr66.66
Ni24Zr76
1533
1283
1283
1393
1233
10.49
11.53
11.56
9.64
12.01
5.58
5.18
5.42
4.39
5.51
7.15
6.90
6.88
8.34
6.95
3.99
3.57
3.73
3.66
3.83
[184]
[184]
[69]
[184]
[184]
v2c
8s
ro3 r
"
1
Q2
64p2 ro3 seo
!#
1 Fs ; q; e
Fig. 45. Schematic description of the detection system used for measuring the amplitude of deformation of a levitated sample.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
(13)
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
29
6000
4000
2000
0
-2000
0.2
0.4
0.6
0.8
-4000
-6000
Time (s)
where
Fs ; q; e
(14)
Q2
16p2 ro3 eo
(15)
and
e2 E2 r o eo
(16)
respectively, with E being the applied electric eld. From Eq. (13), it
can be noticed that the surface tension depends on the sample
radius and density. For density, the data obtained with the imaging
technique (see Section 4.1.1) is substituted in Eq. (13). Also, to
ensure that the measured properties were not distorted by sample
evaporation, the radius variation was monitored in real time using
the imaging technique instead of relying solely on a measurement
of the mass of the sample before and after the experiment.
During the experiments, the video images from a highresolution camera were recorded (Fig. 47). After the experiment,
each value of the radius at each oscillation was obtained by image
analysis. This procedure eliminated the measurement error due to
sample evaporation. Moreover, the aspect ratio of the sample (ratio
between the horizontal and vertical radii) was also calculated to
evaluate the experimental error induced by sample rotation.
While measuring surface tension, the liquid sample was
intentionally rotated to suppress the excitation of oscillation
modes other than the P2 (cos u)-mode. On the other hand, the
characteristic oscillation frequency vc and the decay time t were
dependent on the rotation rate of the liquid sample V, by the
following equations [188,189]:
Dvc 19 V
21 vc
vc
2
"
2 #
1
1
2 V
1
3 vc
t Dt t
(17)
(18)
vc
vc
(19)
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
30
(b)
2320
2300
2280
2260
2240
2220
2200
2400
Tm
2500
2600
2700
2800
(a) 2340
2700
2600
2500
2400
2300
Tm
2200
3300 3400 3500 3600 3700 3800
2800
Temperature (K)
(d)
3000
2800
2600
2400
2200
Tm
2000
3200 3300 3400 3500 3600 3700
Temperature (K)
(c)
Temperature (K)
750
700
650
600
Tm
550
2100 2200 2300 2400 2500 2600
Temperature (K)
Fig. 48. Typical measurements for the surface tension obtained with electrostatic levitation: (a) Ru [170]; (b) W [48]: (c) Os [175]; and (d) Al2O3 [117].
Table 7
Measured surface tension data of metals and semiconductors.
Element
Tm (K)
Surface tension g
g(Tm)
Ce
La
Pr
Nd
Gd
Tb
1072
1191
1204
1294
1586
1629
Y
Ti
1796
1941
Pt
Zr
2041
2128
V
Rh
2183
2237
B
Hf
2360
2504
Ru
2607
Ir
Nb
2719
2750
Mo
Ta
Os
Re
W
2896
3290
3306
3441
3695
Reference
(103 N m1)
dg/dT
(103 N m1 K1)
751
752
777
754
822
893
866
804
1557
1555
1800
1459
1500
1935
1940
1925
1060
1614
1612
2256
2269
2241
1937
1967
2290
2154
2480
2710
2477
0.07
0.056
0.052
0.095
0.097
0.096
0.11
0.05
0.16
0.18
0.14
0.244
0.11
0.27
0.30
0.28
0.067
0.10
0.11
0.24
0.27
0.16
0.20
0.17
0.26
0.21
0.34
0.23
0.31
[156]
[157]
[157]
[157]
[159]
[160]
[100]
[162]
[194]
[195]
[165,166]
[32]
[194]
[196]
[104]
[195]
[197]
[105]
[195]
[170]
[195]
[152]
[194]
[195]
[198]
[199]
[170]
[49]
[48,176,177]
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
31
Table 8
Measured surface tension data of oxides [117,195].
Composition
Tm (K)
Surface tension g
g(Tm) (10
Al2O3
BaTiO3
2327
1893
3
Nm
640
349
4.3. Viscosity
4.3.1. Measurement method
The viscosity was determined by the drop oscillation method by
measuring the decay time of the surface oscillation of the levitated
sample [29]. In this method, a sample was molten and brought to a
selected temperature. During the sample heating, electric currents
were applied to four coils to rotate the sample about the vertical
axis at a rate of around 5 Hz. The liquid sample was intentionally
rotated to suppress the excitation of oscillation modes other than
the P2cos(u)-mode. This amount of sample rotation stabilized the
rotation axis and helped eliminate the measurement error due to
the random rotation of the sample. The rotating magnetic eld was
removed before viscosity measurements were carried out. Then, a
P2cos(u)-mode of drop oscillation was induced in the sample by
superimposing a small sinusoidal electric eld on the levitation
eld. An oscillation detection system, illustrated in Fig. 45,
measured the uctuation of the vertical diameter of the molten
sample with a 4096 Hz sampling frequency. The transient signal
that followed the termination of the excitation eld was already
shown in Fig. 46.
This signal was analyzed using an in-house written LabVIEWTM
program, and the characteristic oscillation frequency vc and decay
time t were obtained. This was done many times for a given
temperature and repeated for several temperatures. Using the decay
time t given by the signal, the viscosity h can be found by [201]
rro2
5t
Reference
)
dg/dT (10
0.082
0.03
[154]. From all the data for the refractory materials, it seems
remarkable that other than their very high melting temperatures,
the magnitude of their properties or the behavior of their
temperature dependence are not signicantly different from those
of lower melting point materials, even when compared with
mercury at 300 K.
The surface tension of some molten oxide materials is
presented in Table 8 together with the temperature coefcient.
1
3
Nm
1
1
)
[117]
[200]
Sometimes, it suppresses the sample oscillation whereas sometimes it promotes the oscillation. This introduces a large
experimental uncertainty in the decay time, and the measured
viscosity values contain large experimental errors. In order to
minimize this effect, it is important to select the sample size and
the position control parameters so that the control frequency does
not overlap with the characteristic oscillation frequencies of the
sample. In early measurements, the position control frequency
range was DC to 360 Hz (720 Hz sampling rate) which overlapped
the sample oscillation frequency (150300 Hz) and introduced
large experimental uncertainties. In recent experiments, the
position control frequency range was reduced from DC to
120 Hz (240 Hz sampling rate). Another improvement was
recently made with regard to oxidization of the sample. It was
noticed during the viscosity measurements of rare earth metals
that the sample surface tended to be oxidized near the melting
temperature which resulted in erroneous data. In order to monitor
the oxygen partial pressure, a quadruple mass spectrometer was
attached to the chamber and conrmed that the oxygen partial
pressure was no larger than 2 106 Pa [50,202]. Also, the surface
of the levitated sample was monitored by a camera to conrm that
the surface was clean and that no sample oxidization occurred.
Samples of several elements (titanium, nickel, zirconium,
niobium, ruthenium, rhodium, hafnium, iridium, platinum, and
terbium) were weighed and melted into spheroids using an arc
melting furnace. The samples were weighed again and samples of
proper size were selected so that the characteristic oscillation
frequencies were above 160 Hz. The purities of the samples, their
mass and the frequency ranges of their characteristic oscillations
are shown in Table 9.
(20)
v2c 8grro3
(21)
The viscosity measurements with the oscillating drop technique assume that the levitated droplet is free from any external
forces. However, in order to levitate a sample against gravity, a
strong electric eld is applied. Moreover, since a high speed
feedback position control is applied, viscosity measurements with
electrostatic levitators are conducted under strong electric elds
with DC and AC components.
The effect of this position control force on the viscosity
measurements was studied and is summarized in Fig. 49 [50,202].
When the frequency range of the sample position control overlaps
with the characteristic oscillation frequency of the molten sample,
the sample positioning force interacts with the sample oscillation.
Fig. 49. Conceptual drawing showing how to avoid the effect of sample positioning
force in the viscosity measurement using the oscillating drop method: (a) relation
between position control frequency and sample oscillation frequency from DC to
360 Hz frequency range, (b) relation between position control frequency and
sample oscillation frequency from DC to 240 Hz [202].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
32
Table 9
Purity, size and typical characteristic oscillation frequencies of the samples
[50,202].
Material
Purity (%)
Characteristic oscillation
frequency (Hz)
Ti
Ni
Zr
Nb
Ru
Rh
Hf
Ir
Pt
Tb
99.9
99.997
99.5
99.9
99.9
99.9
97
99.9
99.98
99.9
1322
3037
3442
2627
3032
3638
3639
2830
3031, 5354
2325
245355
195220
166187
246250
245258
202213
183192
253262
224228, 165166
167174
(b) 15
Viscosity (mPa.s)
Viscosity (mPa.s)
(a) 10
6
4
2
12
Tm
6
3
Tm
0
3300 3400 3500 3600 3700 3800
Temprature (K)
0
1550 1600 1650 1700 1750 1800 1850
Temperature (K)
9
7
5
3
Tm
1
2600 2700 2800 2900 3000 3100
Temperature (K)
Viscosity (mPa.s)
(d) 7
(c)
Viscosity (mPa.s)
6
5
4
3
2
1
0
2300
Tm
2400
2500
Temperature (K)
2600
Fig. 50. Typical measurements for the viscosity obtained with electrostatic levitation: (a) Gb [159]; (b) W [205]; (c) Mo [198]; and (d) B [197].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
33
Table 10
Measured viscosity data of metals and semiconductors.
Element
Viscosity h
Tm (K)
h(Tm) (10
Ce
La
Pr
Nd
Gd
Tb
1072
1191
1204
1294
1586
1629
Y
Ti
1796
1941
Pt
2041
Zr
2128
V
Rh
2183
2237
B
Hf
2360
2504
Ru
2607
Ir
2719
Nb
2750
Mo
Ta
Os
Re
W
2896
3290
3306
3441
3695
a
b
3
Reference
Pa s)
3.31
2.75
2.82
5.07
4.9
5.3
12.0
3.6
3.3
4.4
7.1
4.82
4.7
4.7
4.83
4.3
4.9
2.9
2.8
7.1
5.2
5.0
6.1
6.0
7.0
5.0
4.5
5.6
8.6
4.2
7.9
6.9
Note
h = h0exp(E/RT)
h0 (103 Pa s)
E (103 J/mol)
0.012
0.381
1.38
0.43
1.7
1.399
0.583
0.00287
0.61
0.033
1.72
0.25
0.37
0.76
50.5
19.6
7.14
26.5
14
18.2
41
106.5
27.0
76.6
22.8
49.9
44.8
31.8
1.23
1.31
0.09
[156]
[157]
[157]
[157]
[159]
[193]
[160]
[162]
[193]
[171]
[193]
[166,167]
[193]
[168]
[32]
[194]
[193]
[104]
[197]
[195]
[105]
[195]
[171]
[195]
[152,172]
[195]
[172]
[198]
[199]
[175]
[49]
[205]
22.7
24.7
64.3
0.84
0.50
0.47
0.60
0.59
1.85
1.05
0.55
0.27
0.004
0.0017
0.08
0.11
44.4
48.7
51.2
49.8
52.2
30.0
35.5
48.9
73.0
213
220
133
128
a
a
a
a
b
a
Improved procedure.
It can hardly t the Arrhenius function.
Table 11
Measured viscosity data of oxides.
Composition
Tm (K)
Viscosity h
Reference
h(Tm) (mPa s)
Al2O3
BiTiO3
2327
1893
h = h0exp(E/RT)
30
15
with strong turbulent ows that are induced inside the molten
drop and hinder viscosity measurements on the ground [209,210].
It is therefore obvious that it is a strong advantage of electrostatic
levitation that containerless viscosity measurements are possible
within certain limits [150]. Viscosity data on binary ZrNi alloys
have been published by Ohsaka et al. [184]. Their results can be
represented by an Arrhenius equation with activation energy for
viscous ow E and a pre-exponential factor h1. They are
summarized in the following table (Table 12).
As described in Section 4.3.1, a high resolution-high speed
camera is used in DLR for determination of the time dependent
radius signal Rz(t) after the P2cos(u) mode has been excited. The
camera has a resolution of 480 500 pixels and records images
with a frame rate of 2000 Hz [150]. The use of a high-speed camera
allows to collect 1020 data points per oscillation and to determine
the vertical radius Rz(t) with a precision of better than 1%. For an
entire viscosity measurement, the duration of an experimental run
is roughly 15 min. An example of a measured R(t) curve is shown in
Fig. 51 for a liquid Zr64Ni36 sample at 1182 K. As shown by Becker
h0 (mPa s)
E (kJ/mol)
3.2
0.53
43.2
53.5
[117]
[200]
Table 12
Arrhenius constants for the viscosities of liquid binary ZrNi alloys [184].
Alloy
TL (K)
h (mPa s)
E (kJ/mol)
Ni50Zr50
Ni36Zr64
Ni33.33Zr66.66
Ni24Zr76
1533
1283
1393
1233
0.105
0.0028
0.065
0.0025
63.0
95.8
67.0
90.7
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 51. Measured vertical radius as function of time (symbols) for a liquid Zr64Ni36
sample at 1182 K. The dots represent the experimental data and the solid line
represents a t of a damped sine curve [69].
For T > TL, h assumes values between 11.5 mPa s and 8 mPa s. In
Fig. 52, the data are shown together with the viscosities reported
by Ohsaka et al. [184]. For T < TL, there is an agreement within 5%.
For T > TL, however, the two curves signicantly deviate from each
other. This deviation increases with an increase of temperature. At
1500 K, the viscosity reported by [184] is about 30% smaller. To
exclude the possibility that this discrepancy could be caused by a
systematic error in the data, due to some non-linear effect in the
surface oscillation, the experiment was performed in DLR with ve
samples having masses of 20.8 mg, 27.0 mg, 38.2 mg, 41.0 mg, and
86.2 mg [69].
Such a non-linear effect could originate, for instance, from an
excitation of ow vortices in the bulk of the droplet when the
oscillation amplitude is too large [206]. Or, as described in Section
4.3.1, from an interference of the surface oscillation by the
positioning controller [50] which would also lead to an apparently
increased viscosity. In case that such a non-linear effect existed in
the droplet oscillation, the use of samples with different masses
would result in apparently different values of the viscosity. As
illustrated in Fig. 51, however, the same curve, h(T), is obtained for
Fig. 52. Viscosity of liquid Ni36Zr64 versus temperature. Solid symbols represent
data measured by Brillo et al. [69] and the open symbols are data from Ohsaka et al.
[184]. The dashed line is an Arrhenius-t and the solid line is a VogelFulcher
Tamman (VFT) t.
Fig. 53. Product of Ni self-diffusion coefcient and the viscosity versus temperature
(symbols) [69]. The solid line is a linear t with zero slope and the dashed line shows
the StokesEinstein (SE) relation.
(22)
dt
M
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
35
Table 13
Measured ratio of constant pressure heat capacity over hemispherical total emissivity for some metals.
Metal
Tm (K)
Cp/eT(Tm) (J/mol K)
Reference
Ti
Zr
1941
2128
3.21
V
Rh
Hf
Ru
Ir
Nb
Mo
2183
2237
2504
2607
2719
2750
2896
45.5
132.28
139.4
150.4
184.0
133.1
179.3
159.5
134.4
159.8
[164]
[32]
[168]
[169]
[104]
[105]
[171]
[171]
[172]
[173]
8.49
8.0
3.15
5.5
5.28
(b) 250
(J/mol.K)
190
180
200
Solid
150
Liquid
170
160
150
2100
Cp
Cp
(J/mol.K)
(a) 200
4.5. Emissivity
2300
2500
2700
Temperature (K)
100
Tm
Tm
2900
50
1600
2000
2400
2800
3200
Temperature (K)
Fig. 54. Ratio of constant pressure heat capacity over hemispherical total emissivity for: (a) liquid and supercooled Ru [170] and (b) for solid, liquid, and supercooled Nb [172].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
36
eT
sample, such that the position sensing lasers (532 nm and 645 nm)
and the UV light (385 nm).
After the levitation experiment, the FTIR spectrometer and the
spectrophotometer were oriented toward the blackbody furnace
for calibration (Fig. 55). The blackbody furnace was equipped with
a simulated chamber wall with an observation window. A pinhole
was set in front of the blackbody to mimic a levitated sample. The
temperatures of the blackbody were changed from 1473 K to
1773 K with 50 K interval to get the radiation intensity of
blackbody as a function of temperature. The translation functions
from spectrometer signals to the radiation intensity of light were
calculated using the calibration data obtained with the blackbody.
Then, the data of levitation experiment were converted to the
radiation intensity of the sample Is.
The emissivity is the ratio of the energy radiated by a particular
material to the energy radiated by a blackbody at the same
temperature. The radiation intensity from a blackbody IB depends on
wavelength l, temperature T, and Is depends on l, T, and direction of
el; u; T
Is l; u; T
IB l; T
eRlIB l; Tdl
IB l; Tdl
R l2
l1
elIB l; Tdl
R l2
l1 IB l; Tdl
R l2
l IS l; Tdl
R l1
2
l1 I B l; Tdl
(25)
(23)
where u is the polar angle measured from the normal to the surface
of a sample. The radiation intensity from a blackbody IB is given by
Plancks law of radiation as follows [215]:
IB
2C 1
l5 expC 2 =lT 1
(24)
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
37
rel
t re f rel;re f
t
(29)
t Mr2
dv 8p 5 dv
rr
dt
15
dt
(30)
Eq. (29) implicitly assumes that the sample size and the
distance between the sample and coils are kept constant. However,
it is impossible to make the sample sizes exactly the same as that of
a reference sample. The relations between the sample size and
induced torque could be derived from the work by Braunbek [231].
According to his calculation, the torque t induced on a spherical
sample (radius r, electrical resistivity rel) by a rotating magnetic
eld (intensity H, rotation rate vs) can be calculated from:
t/
p 1
vs r5 H2
4 rel
(31)
dt
rrel
(32)
rre f dv=dtre f
rel
rel;re f
rdv=dt
(33)
where
s
vs v
vs
(27)
(28)
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
am2pRT 1=2
tA
(34)
d1=T
2:303R
(35)
10
Vapor Pressure (Pa)
38
1
4.5
5.5
6.5
Liquid
0.1
Solid
0.01
0.001
10000/T (K-1)
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
T range (K)
Vp(Tm) (Pa)
Liquid
Solid
9.15417,978 T1
16.63432,960 T1
17002050
17701940
0.795
0.469
39
However, if we compare the ESL data for the solid phase with those
of other investigators, the difference could arise from the departure
from a perfect sphere of our solid sample that induced errors in the
digitization procedure. Additional discrepancies could be attributed to the difference in processing techniques and temperature
measurement. The use of electrostatic levitation and radiative
heating in vacuum isolated the samples from container walls and
gases, whereas the cited authors employed methods for which
possible chemical reactions between the highly reactive titanium
and the crucible or the residual surrounding gas could have
occurred. In addition, electron bombardment and induction
heating, used by the conventional instruments, might have
contaminated the specimen under study through evaporation
from the electrodes. The disappearing lament technique utilized
by some authors to measure temperature might also have
introduced some errors. As an additional potential source of
discrepancy, the Knudsen effusion equation was used to calculate
the vapor pressure for electrostatically levitated droplets. However, the physical reality between the geometry of the spheroid
sample and that of normally used samples (plane) was different. To
some extent, the atoms on the surface of the ESL sample had fewer
neighbors that those lying in a plane, which was the case for other
investigators [233,238,239]. Therefore, this could further explain
the lesser energy (or lower vapor pressure) observed in the ESL
experiments.
Electrostatic levitation with an imaging technique permitted
the measurement of the vapor pressure of liquid and supercooled
titanium. A drawback explaining that this method has not been
used for other materials is the coating of the window of the
chamber which required extensive cleaning and maintenance.
However, besides this issue, the method could also, in principle, be
applied to any solid or liquid materials, as long as they can be
levitated. However, as is the case with any other techniques, it
could be time consuming if low vapor pressure materials or if low
temperatures are investigated [232].
5. Short-range structure and atomic dynamics
Fig. 61. Comparison between the values obtained by experiment and calculation
based on the hard sphere model for surface tension (a) and viscosity (b) of hafnium
[250].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
40
Table 15
Comparison of hard sphere calculation and experimental values at the melting temperature for heat capacity (Cp), surface tension (g), and viscosity (h) [250].
Element
Ti
Zr
Nb
Hf
34.2
33.8
33.4
33.8
Viscosity h (103 Pa s)
Experiment
Calculation
Experiment
Calculation
Experiment
35.9
41.7
40.3
34.3
1634
1445
2177
1646
1557
1500
1937
1614
3.0
3.6
4.7
5.4
4.4
4.7
4.5
5.2
Fig. 62. Illustration of the main system components, as viewed from the top, of the
JAXA vacuum electrostatic levitator dedicated to neutron beam interaction studies.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
41
Fig. 63. Side view of the JAXA neutron vacuum electrostatic levitator.
Fig. 64. Intensity of alumina and background with the Cd slit: (a) alumina, (b)
background [58].
Reprinted with permission from H. Aoki et al., Rev. Sci. Instrum. 74, 11471149.
Copyright 2003, AIP Publishing LLC.
Fig. 65. Intensity of alumina with the corresponding mirror indices [58].
Reprinted with permission from H. Aoki et al., Rev. Sci. Instrum. 74, 11471149.
Copyright 2003, AIP Publishing LLC.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 66. Schematic drawing (vertical cut) of the DLR electrostatic levitator for
neutron scattering experiments. Sample: S; top electrode: TE; ground electrode:
GE; side electrodes: SE; high voltage feed through: HV; positioning laser with beam
expander: L; mirror: M3; position sensitive detector: PSD; neutron window: NW;
turbo molecular pump: TP; UV lamp: UV [261].
Fig. 67. Schematic drawing (horizontal cut) of the electrostatic levitator integrated
in the neutron scattering environment. Neutron beam guide: NG; neutron beam:
NB; neutron slit system: NS; neutron B4C aperture: NA; neutron window: NW;
neutron beam stop: BS; neutron detector: D; mirrors: M1-3; position sensitive
detector: PSD; high voltage feedthrough: HV; side electrodes: SE; UV lamp: UV;
laser head containing the coupling optics of the heating laser, the pyrometer, and
the CCD video camera: LH; window for sample exchange: W [261].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
Fig. 68. Raw diffractogram of liquid Zr36Ni64 (0.6 g mass) at a temperature of 1370 K
measured using a neutron wavelength of 0.94 A. Also shown is the background of
the empty ESL, which is one order of magnitude smaller than that of the signal of the
sample.
43
Fig. 69. Total static structure factor of liquid Zr64Ni36 for different temperatures in
the stable liquid and in the regime of the supercooled melt below the liquidus
temperature of TL = 1283 K. The structure factors are shifted by multiples of 0.5 for
better visibility [261].
Fig. 70. Total structure factors measured by neutron scattering and X-ray diffraction
for liquid Zr50Cu50 at different temperatures above and below the liquidus
temperature of TL = 1226 K [262].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 71. Partial structure factors SNN (upper curves) and SNC (lower curves) for liquid
Zr50Cu50 at T = 1280 K (red) and T = 1175 K (blue) [262] (For interpretation of the
references to color in this gure legend, the reader is referred to the web version of
the article.).
Fig. 72. Raw time-of-ight spectrum measured for a Zr64Ni36 melt at T = 1478 K
acquired within a range of scattering angles of 598 < 2u < 748. Also shown is a
spectrum of the empty levitator.
Fig. 73. Scattering law S(q,v) of liquid Zr64Ni36 for q = 0.5 A1 at T = 1115 K (lled
symbols) and T = 1460 K (open symbols) tted with Lorentzian proles [272].
Table 16
Nearest neighbor (r1) and second neighbor (r2) distances, and nearest neighbor
coordination numbers, ZNN, for ZrCu melts at different temperatures [262].
Zr66.7Cu33.3
Zr50Cu50
Zr41.2Cu58.8
T (K)
r1 (A)
r2 (A)
ZNN
1325
1175
1280
1110
1260
2.95 0.01
2.79 0.01
2.79 0.01
2.75 0.01
2.74 0.01
5.37 0.05
5.10 0.05
5.13 0.05
5.01 0.05
4.94 0.05
13.7 0.5
13.8 0.5
13.6 0.5
13.8 0.5
13.6 0.5
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
45
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 76. Difference between the Compton proles for the solid and liquid phases,
[DJ(pz) = Jsolid(pz) Jliquid(pz)]: experiment (black) and CPMD simulation (red) [61]
(For interpretation of the references to color in this gure legend, the reader is
referred to the web version of the article.).
Fig. 77. Snapshot from the simulation on l-Si at 1787 K illustrating Si atoms
(yellow), covalent bond pairs (green), lone pairs (red), and diffuse pairs (translucent
blue). Note that bonds connecting Si atoms are only guide for the eyes [61] (For
interpretation of the references to color in this gure legend, the reader is referred
to the web version of the article.).
Fig. 78. Calculated structure factors of possible local cluster types in liquid: (a)
tetrahedron, (b) bi-tetrahedron (the dotted line is for a distorted tetrahedron), (c)
icosahedron, and (d) decahedron are energetically favored structures as well as (f);
(e) the Z14 polyhedron in FrankKasper phases is a prominent cluster type in
glasses [127].
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 77, 184102. Copyright
2008, American Physical Society.
was suggested for the liquid structure [280]. However, this model
did not agree with the structure of early transition liquid metals
obtained in experiments [281]. The disagreement may be caused
by either modeling or experiments. From the experimental view,
the combination of electrostatic levitation and high energy
synchrotron X-ray scattering can give accurate structural information of the liquid metals, since the scattered intensity from a
levitated liquid does not need correction by a container holding the
liquid sample, and the transmitted scattering can give the bulk
structural information of the liquid metals. The characteristic
feature of the local order of the liquid appears in high q range with
high energy X-ray scattering experiment as seen in Fig. 78.
In Fig. 79, the structure factor (S(q)) of liquid metals (Ni, Ti, and
Zr) is shown. The calculated structure factors cannot t the
Fig. 79. Experimentally measured structure factors for pure elemental liquids (Ni,
Ti, and Zr) in the equilibrium and supercooled states. The dashed lines indicate an
optimized t to the PercusYevick hard sphere model [127].
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 77, 184102. Copyright
2008, American Physical Society.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
47
Fig. 81. Comparison of the experimental data (open circles) and the calculated S(q)
of Ti with ISRO (thin line) and DISRO (thick line) including DebyeWaller (DW)
factor. The inset shows the calculated S(q) for different degrees of distortion
without DW factor; the solid line is calculated for the above distortions, and the
dotted and thin lines are for D3d (threefold) and D2d (twofold) [282] distortions,
respectively [126].
Reprinted with permission from G.W. Lee et al., Phys. Rev. Lett. 93, 037802.
Copyright 2004, American Physical Society.
Fig. 80. X-ray structure factors for liquid Ni (liquidus temperature, TL = 1728 K) (a)
and Ti (TL = 1940 K) (b) as a function of temperature (connected open circles); the
curves are displaced vertically for clarity. Dashed lines are from the PercusYevick
hard sphere model. Selected (for clarity) neutron diffraction data from EML
experiments [128] for Ni are shown as closed circles in (a) and the X-ray data from
earlier work on Ti [286] are shown by the dotted line in (b). Insets show the t of the
experimental data (connected open circles) to a 13-atom icosahedron (solid lines),
above and below the melting temperatures [126].
Fig. 82. (a) Calculated S(q) for supercooled liquid Ti, short-bond length atoms (SRO
clusters, black line), and background atoms with long bonds (gray line); (b) bond
angle distributions [285].
Reprinted with permission from B. Lee et al., J. Chem. Phys. 129, 024711. Copyright
2008, AIP Publishing LLC.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
1=N
i; j
!
2
f i f j sinqr i; j =qri; j h f i
2
h fi 1
(36)
Fig. 83. (a) Structure factors and (b) reduced radial distribution functions for liquid
Si, obtained from diffraction patterns taken at a frame rate of 1 Hz during cooling at
13 K/s [287].
Reprinted with permission from T.H. Kim et al., Phys. Rev. Lett. 95, 085501.
Copyright 2005, American Physical Society.
Fig. 84. The coordination number of supercooled liquid Si (left). In order to show the
change in coordination during recalescence, a time series is displayed in the inset.
The arrow denotes the onset of recalescence. The data were taken with 10 Hz frame
rate [287]. Temperature dependence of the coordination number obtained by
previous measurements (* [244]), & [288], ~ [289]).
Reprinted with permission from T.H. Kim et al., Phys. Rev. Lett. 95, 085501.
Copyright 2005, American Physical Society.
Fig. 86. (a) A 13-atom icosahedral cluster showing the bond types considered for the
calculation of S(q); (b) total S(q) and the contributions of each bond type to S(q); (c)
the combinations of the various bond types to S(q) [127].
Reprinted with permission from G.W. Lee et al., Phys. Rev. B 77, 184102. Copyright
2008, American Physical Society.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
For the ISRO with 13 atoms and with four type lengths, S(q) is given
as follows.
1=1324 f 1 f 2 sinqr1=qr1 60 f22 sinqr2=qr2
49
Sq
h fi 1
(37)
The partial S(q) for each bond length is shown in Fig. 86(a).
Among the bond lengths, r2 and r3 type-bonds dominantly affect
the second peak shape of the total S(q). Therefore, when the
number of r2 and r3 bonds increases, the behavior of the second
oscillation on S(q) can be reproduced with supercooling in
Fig. 86(b). The r2 and r3 type-bonds are found in the distorted
tetrahedral order in Fig. 86(a), which is the unit structure for
ISRO. Therefore, it can be concluded that the fragments of ISRO,
i.e., distorted tetrahedral order, increase with supercooling.
Fig. 89. Calculated S(q) for Ti40Zr30Ni30 as a function of distortion angle (a), and a
comparison of the calculated S(q) for different degrees of distortion (b). The thick
line was computed for a combination of perfect icosahedra (55%) and distorted
icosahedra with u = 78 (45%). The arrows indicate the position of the third and fourth
peaks [136].
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
Fig. 90. Static structure factor for (a) liquid Zr80Pt20 at 1607 8C, 1474 8C, 1345 8C,
1217 8C, 1090 8C, and 978 8C. The insets show the evolution of S(q) with decreasing
temperature for (b) the primary peak, (c) the second peak, and (d) the pre-peak
(indicated with an arrow in the total structure factor) [296].
Reprinted gure with permission from N.A. Mauro et al., Phys. Rev. B 83, 184109 (8
pp.), 2011. Copyright 2011 by the American Physical Society.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
[82]
[83]
[84]
[85]
[86]
51
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
P.-F. Paradis et al. / Materials Science and Engineering R xxx (2013) xxxxxx
[87] G.W. Lee, A.K. Gangopadhyay, T.K. Croat, T.J. Rathz, R.W. Hyers, J.R. Rogers, K.F.
Kelton, Phys. Rev. B 72 (2005) 174107.
[88] K.F. Kelton, G.W. Lee, A.K. Gangopadhyay, R.W. Hyers, T.J. Rathz, J.R. Rogers, M.B.
Robinson, D.S. Robinson, Phys. Rev. Lett. 90 (2003) 195504 (3 pp.).
[89] F.C. Frank, Proc. R. Soc. Lond. A 215 (1952) 4346.
[90] G.W. Lee, A.K. Gangopadhyay, K.F. Kelton, Acta Mater. 59 (2011) 49644973.
[91] D.R. Nelson, F. Spaepen, in: HEAD Turnbull (Ed.), Solid State Physics, vol. 42,
Academic Press, Boston, MA, 1989.
[92] J.F. Lofer, J. Schroers, W.L. Johnson, Appl. Phys. Lett. 77 (2000) 681683.
[93] C.C. Hays, C.P. Kim, W.L. Johnson, Appl. Phys. Lett. 75 (1999) 10891091.
[94] C.C. Hays, J. Schroers, W.L. Johnson, T.J. Rathz, R.W. Hyers, J.R. Rogers, M.B.
Robinson, Appl. Phys. Lett. 79 (2001) 16051607.
[95] T.H. Kim, A.K. Gangopadhyay, L.Q. Xing, G.W. Lee, Y.T. Shen, K.F. Kelton, A.I.
Goldman, R.W. Hyers, J.R. Rogers, Appl. Phys. Lett. 87 (2005) 251924 (3 pp.).
[96] J.K.R. Weber, J. Tangeman, T. Key, K. Hiera, P.-F. Paradis, T. Ishikawa, J. Yu, S. Yoda,
Jpn. J. Appl. Phys. 41 (Pt. 1 (5A)) (2002) 30293030.
[97] P.-F. Paradis, J. Yu, T. Ishikawa, S. Yoda, J. Am. Ceram. Soc. 86 (2003) 22342236.
[98] P.-F. Paradis, J. Yu, T. Ishikawa, T. Aoyama, S. Yoda, J.K.R. Weber, J. Cryst. Growth
249 (2003) 523530.
[99] K.S. Ranasinghe, C.S. Ray, D.E. Day, J.R. Rogers, R.W. Hyers, T. Rathz, J. Mater. Sci.
42 (2007) 42914297.
[100] T. Ishikawa, P.-F. Paradis, T. Itami, S. Yoda, Meas. Sci. Technol. 16 (2005) 443
451.
[101] Y.S. Sung, H. Takeya, K. Togano, Rev. Sci. Instrum. 72 (2001) 44194423.
[102] Y.S. Sung, H. Takeya, K. Hirata, K. Togano, Appl. Phys. Lett. 82 (2003) 36383640.
[103] Y.S. Sung, H. Takeya, K. Togano, J. Appl. Phys. 93 (2003) 36813683.
[104] P.-F. Paradis, T. Ishikawa, S. Yoda, Int. J. Thermophys. 24 (2003) 11211135.
[105] P.-F. Paradis, T. Ishikawa, S. Yoda, Int. J. Thermophys. 24 (2003) 239257.
[106] P.-F. Paradis, T. Ishikawa, S. Yoda, Appl. Phys. Lett. 83 (2003) 40474049.
[107] K.F. Kelton, A.K. Gangopadhyay, G.W. Lee, L. Hennet, R.W. Hyers, S. Krishnan,
M.B. Robinson, J. Rogers, T.J. Rathz, J. Non-Cryst. Solids 312314 (2002) 305308.
[108] M.B. Robinson, D. Li, J.R. Rogers, R.W. Hyers, L. Savage, T.J. Rathz, Appl. Phys. Lett.
77 (2000) 32663268.
[109] T.J. Rathz, M.B. Robinson, R.W. Hyers, J.R. Rogers, D. Li, J. Mater. Sci. Lett. 21
(2002) 301303.
[110] D.M. Matson, D.J. Fair, R.W. Hyers, J.R. Rogers, Ann. N. Y. Acad. Sci. 1027 (2004)
435446.
[111] T. Aoyama, P.-F. Paradis, T. Ishikawa, S. Yoda, Mater. Sci. Eng. A 375377 (2004)
460463.
[112] P.-F. Paradis, T. Ishikawa, J. Yu, unpublished results, 2003.
[113] T. Matsushita, T. Ishikawa, P.-F. Paradis, K. Mukai, S. Seetharaman, Iron Steel Inst.
Jpn. Intl. 46 (2006) 606610.
[114] I. Minato, H. Fukayama, T. Ishikawa, P.-F. Paradis, J. Yu, S. Yoda, Int. J. Thermophys. 27 (2006) 11731180.
[115] P.-F. Paradis, J. Yu, T. Ishikawa, T. Aoyama, S. Yoda, Appl. Phys. A 79 (2004) 1965
1969.
[116] P.-F. Paradis, T. Ishikawa, Y. Saita, S. Yoda, Jpn. J. Appl. Phys. 43 (4A) (2004) 1496
1500.
[117] P.-F. Paradis, T. Ishikawa, Jpn. J. Appl. Phys. 44 (7A) (2005) 50825085.
[118] J. Yu, P.-F. Paradis, T. Ishikawa, S. Ozawa, T. Saito, T. Motegi, S. Yoda, Jpn. J. Appl.
Phys. 41 (2002) 29082911.
[119] J. Yu, P.-F. Paradis, T. Ishikawa, S. Yoda, J. Korean Phys. Soc. 46 (2005) 1518.
[120] J. Yu, P.-F. Paradis, T. Ishikawa, S. Yoda, Ferroelectrics 301 (2004) 199201.
[121] J. Yu, P.-F. Paradis, T. Ishikawa, S. Yoda, Jpn. J. Appl. Phys. 43 (2004) 81358138.
[122] J. Yu, P.-F. Paradis, T. Ishikawa, S. Yoda, Appl. Phys. Lett. 85 (2004) 28992901.
[123] J. Yu, P.-F. Paradis, T. Ishikawa, S. Yoda, Y. Saita, M. Itoh, F. Kano, Chem. Mater. 16
(2004) 39733975.
[124] J. Yu, P.-F. Paradis, T. Ishikawa, I. Miura, S. Yoda, Y. Shan, J. Cryst. Growth 273
(2005) 515519.
[125] K.F. Kelton, in: H. Ehrenreich, D. Turnbull (Eds.), Solid State Physics, Vol. 45,
Academic Press, Boston, 1991, p. 75.
[126] G.W. Lee, A.K. Gangopadhyay, K.F. Kelton, R.W. Hyers, T.J. Rathz, J.R. Rogers, D.S.
Robinson, Phys. Rev. Lett. 93 (2004) 037802 (3 pp.).
[127] G.W. Lee, A.K. Gangopadhyay, R.W. Hyers, T.J. Rathz, J.R. Rogers, D.S. Robinson,
A.I. Goldman, K.F. Kelton, Phys. Rev. B 77 (2008) 184102 (8 pp.).
[128] T. Schenk, D. Holland-Moritz, V. Simonet, R. Bellissent, D.M. Herlach, Phys. Rev.
Lett. 89 (2002) 075507 (3 pp.).
[129] V. Simonet, F. Hippert, H. Klein, M. Audier, R. Bellissent, H. Fischer, A.P. Murani,
D. Boursier, Phys. Rev. B 58 (1998) 62736286.
[130] D. Holland-Moritz, Int. J. Non-Equilib. Process. 11 (1998) 169199.
[131] G.W. Lee, A.K. Gangopadhyay, K.F. Kelton, J. Alloys Compd. 537 (2012) 171174.
[132] V.P. Skripov, in: E. Kaldis, H. Scheel (Eds.), Current Topics in Materials and
Science, Crystal Growth and Materials, vol. 2, North Holland, Amsterdam, 1977,
pp. 327378.
[133] C.W. Morton, W.H. Hofmeister, R.J. Bayuzick, A.L. Rulison, J.L. Watkins, Acta
Mater. 46 (1998) 60336039.
[134] W.H. Hofmeister, C.W. Morton, R.J. Bayuzick, Acta Mater. 46 (1998) 19031908.
[135] D. Turnbull, Contemp. Phys. 10 (1969) 473488.
[136] G.W. Lee, Undercooling, local structure and phase transition in icosahedral
quasicrystals forming titanium-zirconium-nickel alloy liquids, Ph.D. Thesis,
Washington University, 2004.
[137] N. Mattern, J.H. Han, M. Zinkevich, O. Shuleshova, I. Kaban, D. Holland-Moritz, J.
Gegner, F. Yang, J. Bednarcik, W. Loser, J. Eckert, Calphad 39 (2012) 2732.
[138] T. Iida, R.I.L. Guthrie, The Physical Properties of Liquid Metals, Clarendon Press,
Oxford, 1988.
[139] D. Jones, Nature 398 (1999) 565566.
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001
G Model
53
Please cite this article in press as: P.-F. Paradis, et al., Mater. Sci. Eng. R (2013), http://dx.doi.org/10.1016/j.mser.2013.12.001